torchlm 0.1.6.1

A PyTorch landmarks-only library with 100+ data augmentations, training and inference, can easily install with pip and compatible with albumentations and torchvision.

English | 中文文档

🤗 Introduction

torchlm is aims to build a high level pipeline for face landmarks detection, it supports training, inference and 100+ data augmentations, can easily install with pip.

❤️ Star 🌟👆🏻 this repo to support me if it does any helps to you, thanks ~

👋 Core Features

• High level pipeline for training and inference.
• Provides 30+ native landmarks data augmentations.
• Can bind 80+ transforms from torchvision and albumentations with one-line-code.
• Support awesome models for face landmarks detection, such as YOLOX, YOLOv5, ResNet, MobileNet, ShuffleNet and PIPNet, etc.

🆕 What's New

• [2022/03/08]: Add PIPNet: Towards Efficient Facial Landmark Detection in the Wild, CVPR2021
• [2022/02/13]: Add 30+ native data augmentations and bind 80+ transforms from torchvision and albumentations.

🛠️ Usage

Requirements

• opencv-python-headless>=4.5.2
• numpy>=1.14.4
• torch>=1.6.0
• torchvision>=0.8.0
• albumentations>=1.1.0
• onnx>=1.8.0
• onnxruntime>=1.7.0
• tqdm>=4.10.0

Installation

you can install torchlm directly from pypi. See NOTE before installation!!!

pip3 install torchlm
# install from specific pypi mirrors use '-i'
pip3 install torchlm -i https://pypi.org/simple/

or install from source if you want the latest torchlm and install it in editable mode with -e.

# clone torchlm repository locally if you want the latest torchlm
git clone --depth=1 https://github.com/DefTruth/torchlm.git 
cd torchlm
# install in editable mode
pip install -e .

NOTE: If you have the conflict problem between different installed version of opencv (opencv-python and opencv-python-headless, ablumentations need opencv-python-headless). Please uninstall the opencv-python and opencv-python-headless first, and then reinstall torchlm. See albumentations#1139 for more details.

# first uninstall confilct opencvs
pip uninstall opencv-python
pip uninstall opencv-python-headless
pip uninstall torchlm  # if you have installed torchlm
# then reinstall torchlm
pip install torchlm # will also install deps, e.g opencv

🌟🌟Data Augmentation

torchlm provides 30+ native data augmentations for landmarks and can bind with 80+ transforms from torchvision and albumentations through torchlm.bind method. Further, torchlm.bind provide a prob param at bind-level to force any transform or callable be a random-style augmentation. The data augmentations in torchlm are safe and simplest. Any transform operations at runtime cause landmarks outside will be auto dropped to keep the number of landmarks unchanged. The layout format of landmarks is xy with shape (N, 2), N denotes the number of the input landmarks.

• use almost 30+ native transforms from torchlm directly

import torchlm
transform = torchlm.LandmarksCompose([
    torchlm.LandmarksRandomScale(prob=0.5),
    torchlm.LandmarksRandomMask(prob=0.5),
    torchlm.LandmarksRandomBlur(kernel_range=(5, 25), prob=0.5),
    torchlm.LandmarksRandomBrightness(prob=0.),
    torchlm.LandmarksRandomRotate(40, prob=0.5, bins=8),
    torchlm.LandmarksRandomCenterCrop((0.5, 1.0), (0.5, 1.0), prob=0.5)
])

Also, a user-friendly API build_default_transform is available to build a default transform pipeline.

transform = torchlm.build_default_transform(
    input_size=(input_size, input_size),
    mean=[0.485, 0.456, 0.406],
    std=[0.229, 0.224, 0.225],
    force_norm_before_mean_std=True,  # img/=255. first
    rotate=30,
    keep_aspect=False,
    to_tensor=False
)

• bind 80+ torchvision and albumentations's transforms through torchlm.bind

transform = torchlm.LandmarksCompose([
    torchlm.bind(torchvision.transforms.GaussianBlur(kernel_size=(5, 25)), prob=0.5),  
    torchlm.bind(albumentations.ColorJitter(p=0.5))
])

See transforms.md for supported transforms sets and more example can be found at test/transforms.py.

bind custom callable array or Tensor functions through torchlm.bind

# First, defined your custom functions
def callable_array_noop(img: np.ndarray, landmarks: np.ndarray) -> Tuple[np.ndarray, np.ndarray]: # do some transform here ...
    return img.astype(np.uint32), landmarks.astype(np.float32)

def callable_tensor_noop(img: Tensor, landmarks: Tensor) -> Tuple[Tensor, Tensor]: # do some transform here ...
    return img, landmarks

# Then, bind your functions and put it into the transforms pipeline.
transform = torchlm.LandmarksCompose([
        torchlm.bind(callable_array_noop, bind_type=torchlm.BindEnum.Callable_Array),
        torchlm.bind(callable_tensor_noop, bind_type=torchlm.BindEnum.Callable_Tensor, prob=0.5)
])

some global debug setting for torchlm's transform

• setup logging mode as True globally might help you figure out the runtime details

# some global setting
torchlm.set_transforms_debug(True)
torchlm.set_transforms_logging(True)
torchlm.set_autodtype_logging(True)

some detail information will show you at each runtime, the infos might look like

LandmarksRandomScale() AutoDtype Info: AutoDtypeEnum.Array_InOut
LandmarksRandomScale() Execution Flag: False
BindTorchVisionTransform(GaussianBlur())() AutoDtype Info: AutoDtypeEnum.Tensor_InOut
BindTorchVisionTransform(GaussianBlur())() Execution Flag: True
BindAlbumentationsTransform(ColorJitter())() AutoDtype Info: AutoDtypeEnum.Array_InOut
BindAlbumentationsTransform(ColorJitter())() Execution Flag: True
BindTensorCallable(callable_tensor_noop())() AutoDtype Info: AutoDtypeEnum.Tensor_InOut
BindTensorCallable(callable_tensor_noop())() Execution Flag: False
Error at LandmarksRandomTranslate() Skip, Flag: False Error Info: LandmarksRandomTranslate() have 98 input landmarks, but got 96 output landmarks!
LandmarksRandomTranslate() Execution Flag: False

• Execution Flag: True means current transform was executed successful, False means it was not executed because of the random probability or some Runtime Exceptions(torchlm will should the error infos if debug mode is True).

• AutoDtype Info:

  • Array_InOut means current transform need a np.ndnarray as input and then output a np.ndarray.
  • Tensor_InOut means current transform need a torch Tensor as input and then output a torch Tensor.
  • Array_In means current transform needs a np.ndarray input and then output a torch Tensor.
  • Tensor_In means current transform needs a torch Tensor input and then output a np.ndarray.

  But, is ok if you pass a Tensor to a np.ndarray-like transform, torchlm will automatically be compatible with different data types and then wrap it back to the original type through a autodtype wrapper.

🎉🎉Training

In torchlm, each model have a high level and user-friendly API named training, here is a example of PIPNet.

from torchlm.models import pipnet

model = pipnet(
        backbone="resnet18",
        pretrained=False,
        num_nb=10,
        num_lms=98,
        net_stride=32,
        input_size=256,
        meanface_type="wflw",
        backbone_pretrained=True,
        map_location="cuda",
        checkpoint=None
)

model.training(
        self,
        annotation_path: str,
        criterion_cls: nn.Module = nn.MSELoss(),
        criterion_reg: nn.Module = nn.L1Loss(),
        learning_rate: float = 0.0001,
        cls_loss_weight: float = 10.,
        reg_loss_weight: float = 1.,
        num_nb: int = 10,
        num_epochs: int = 60,
        save_dir: Optional[str] = "./save",
        save_interval: Optional[int] = 10,
        save_prefix: Optional[str] = "",
        decay_steps: Optional[List[int]] = (30, 50),
        decay_gamma: Optional[float] = 0.1,
        device: Optional[Union[str, torch.device]] = "cuda",
        transform: Optional[transforms.LandmarksCompose] = None,
        coordinates_already_normalized: Optional[bool] = False,
        **kwargs: Any  # params for DataLoader
) -> nn.Module:

How to train PIPNet using your custom dataset and meanface setting?

• setup your custom meanface and nearest-neighbor landmarks through set_custom_meanface method, this method will calculate the distance between landmarks in meanface and auto setup the nearest-neighbors for each landmarks.

def set_custom_meanface(
        self,
        custom_meanface_file_or_string: str
) -> bool:
    """
    :param custom_meanface_file_or_string: a long string or a file contains normalized
    or un-normalized meanface coords, the format is "x0,y0,x1,y1,x2,y2,...,xn-1,yn-1".
    :return: status, True if successful.
    """

Please jump to the entry point of the function for the detail documentations of training API for each defined models in torchlm, e.g pipnet/_impls.py#L159. Further, the model implementation plan is as follows:

PIPNet	YOLOX	YOLOv5	NanoDet	ResNet	MobileNet	ShuffleNet	VIT	...
✅	❔	❔	❔	❔	❔	❔	❔	❔

✅ = known work and official supported, ❔ = in my plan, but not coming soon.

👀👇 Inference

C++ API

The ONNXRuntime(CPU/GPU), MNN, NCNN and TNN C++ inference of torchlm will be release at lite.ai.toolkit.

Python API

In torchlm, a high level API named runtime.bind can bind face detection and landmarks models together in torchlm and then you can run the runtime.forward API to get the output landmarks and bboxes, here is a example of PIPNet.

import cv2
import torchlm
from torchlm.tools import faceboxesv2
from torchlm.models import pipnet

def test_pipnet_runtime():
    img_path = "./1.jpg"
    save_path = "./1.jpg"
    checkpoint = "./pipnet_resnet18_10x98x32x256_wflw.pth"
    image = cv2.imread(img_path)

    torchlm.runtime.bind(faceboxesv2())
    torchlm.runtime.bind(
        pipnet(
            backbone="resnet18",
            pretrained=True,
            num_nb=10,
            num_lms=98,
            net_stride=32,
            input_size=256,
            meanface_type="wflw",
            backbone_pretrained=True,
            map_location="cpu",
            checkpoint=checkpoint
        )
    )
    landmarks, bboxes = torchlm.runtime.forward(image)
    image = torchlm.utils.draw_bboxes(image, bboxes=bboxes)
    image = torchlm.utils.draw_landmarks(image, landmarks=landmarks)
    cv2.imwrite(save_path, image)

📖 Documentations

• Data Augmentation's API

🎓 License

The code of torchlm is released under the MIT License.

❤️ Contribution

Please consider ⭐ this repo if you like it, as it is the simplest way to support me.

👋 Acknowledgement

The implementation of torchlm's transforms borrow the code from Paperspace .